A Flexible Material Database for Computational Science and Engineering

Simulation tools in the area of computational science and engineering usually require access to a plethora of material parameters for modeling physical phenomena. Academic software projects typ-ically focus on a single approach for providing this capability. For instance, an extensible markup language (XML)-based input file is used, containing all relevant material parameters, such as the per-mittivity for silicon [1]. These parameters are then imported and accessed via XML libraries, like pugixml [2]. Although this approach is effective, it confines the application to use a single specific material database backend. For small simulation projects, this is hardly an issue. However, the larger the project and the higher the required usability, the more it becomes important to support different – typically file-based – material databases and to reuse historically-grown material database files. We tackle these challenges by using a flexible material database mechanism. Flexibility refers to decou-pling a simulation tool from the actual material database kernel, such as XML. In particular, we strive for exchangeability and expandability via a variable database mechanism, based on a dynamic class hierarchy. We ensure proper handling of physical units via incorporating unit checks, which are es-pecially important for robust numerical simulations [3]. Also, we investigate interfacing with Python